关键词: 推荐系统, 图神经网络, 数据增强, 对比学习, 冷启动问题

Abstract: With the exponential growth of information, recommendation systems have played a crucial role in improving user experience and alleviating information overload. However, traditional recommendation methods still face limitations in addressing data sparsity, cold-start issues, and model interpretability. To tackle these challenges, this study proposes a graph contrastive learning convolutional neural network recommendation system based on global graph data augmentation.First, global augmentation is performed at the node, edge, and feature levels to enhance the model’s robustness under sparse data conditions. Second, a multi-granularity graph contrastive learning framework is designed, which conducts contrastive learning tasks at the node, subgraph, and global graph levels to strengthen the consistency and discriminability of the learned embeddings.Experiments conducted on several public datasets, including MovieLens-1M and Amazon-Books, demonstrate that the proposed model outperforms traditional methods such as Matrix Factorization and LightGCN in terms of recommendation accuracy, robustness, and performance in cold-start scenarios—particularly achieving superior results on sparse datasets. Furthermore, ablation studies and parameter sensitivity analyses verify the effectiveness of the global augmentation strategy and the multi-granularity contrastive learning mechanism. These findings indicate that the proposed model shows significant advantages in addressing key challenges in recommendation systems, such as data sparsity and the cold-start problem.

Key words: recommender system, graph neural network, data augmentation, contrastive learning, cold-start problem